Scientist Stories: Arlene Sharpe, Immunoregulation and T-cell Tolerance

I'm delighted to be here this morning. And what I'll be talking about today is the regulation of T lymphocytes. You just heard a beautiful introduction to the immune system and to what T lymphocytes recognize. And I'll talk about essentially mechanisms that turn the immune system on, these T cells on, and also what turns them off. So when we think about an immune response, we have a limited number of T cells that are born that can recognize any specific antigen. So if you get an infection, for example, there is a rapid expansion of these lymphocytes. Studies from viral infections have shown that an antigen-specific T cell can expand 50,000-fold in one week in response to a virus. And so this means that the immune response must be very carefully controlled because of this intense ability to respond to a dangerous signal. So we know that when there is a virus-specific stimulus, for example, that these cells can respond, proliferate, and expand, and then they can differentiate into effector cells to defend us against this microbe. The T cells can differentiate into the helper T cells that can help B cells make antibodies and can enable cytotoxic T cells to differentiate that then can kill the virally infected cells. After the pathogen is gone, there is a decline in the immune response referred to as homeostasis. And then a transition to memory. As you've already heard, one of the special features of the immune system is that it can remember its encounter with the antigen. So the next time it sees a virus, it will respond more effectively and more quickly. This is the basis for vaccination, but also the hope for cancer immunotherapy to have durable responses to the cancer, so it is gone. So we have these passive mechanisms. The responses to pathogens are gone when the pathogen is eliminated. But there are also active mechanisms that balance activation and control. And that's what we're going to be talking about this morning, that there's a balance in the immune system. You want to have activation of the immune system to defend us against the microbial world and tumors, but on the other hand, we don't want to respond to ourselves. We tolerate our own antigens, a system known as tolerance. So we have this balance so that we get the appropriate activation. You sometimes can have too much of a good thing. Excessive T lymphocyte activation can lead to immune-mediated tissue damage, and so we need to regulate the immune responses to protect against this. In addition, there are certain scenarios where there are persistent antigens. The— unlike a microbe that can be cleared and goes away, we constantly see the immune system sees self-antigens Tumors are constantly there, and some chronic infections such as HIV present a chronic antigenic stimulus. So there's a balance that occurs here, and we'll talk about what changes in the immune system when a chronic stimulus is encountered. And failure of these control mechanisms can lead to immune-mediated diseases. So, we're going to talk about T-cell co-stimulation, which is a critical regulator of this balance between T-cell activation and control that helps us respond appropriately to defend against microbial pathogens, but not respond to self. And of course, the breakdown of tolerance leads to autoimmunity. So, both sides of this balance need to be regulated. So first, let's talk about T cell activation and co-stimulation. There is a two-signal model for T cell activation that according to this model, for T cells to become activated, they need to receive two signals that are delivered by these antigen-presenting cells, such as the dendritic cells that you just heard about. The first signal is what confers specificity to an immune response and involves that antigen recognition where antigenic peptides can be presented by the MHC molecules to receptors on the T cells of the T cell receptor. In addition to this signal 1, there is a co-signal or co-stimulatory signal that is also delivered by antigen-presenting cells to T cells. And when the T cells receive both of these signals, they then can become fully activated and can differentiate. They can proliferate, differentiate to become these effector cells. Now, this is a field that's had its origins in actually understanding how B cells became regulated in the 1980s. And then in the late 1980s and early 1990s, this was a hypothesis that's been proven. And so there was great interest in understanding these— what are these co-stimulatory pathways from a fundamental viewpoint, because if we understood these pathways, we could understand a lot about mechanisms that regulate the immune response, but also from the earliest times it was appreciated that if we could understand these co-stimulatory signals, these would be of great therapeutic interest. The idea being that if we understood these signals, we could understand how tolerance is regulated and use this information therapeutically to enable transplantation acceptance, prevent transplant rejection, and also to treat autoimmune diseases that develop when there's a breakdown of tolerance. On the other hand, there are situations where we want to stimulate the immune system. So understanding these co-stimulatory signals, could this be used to develop better vaccines for infectious diseases and used to stimulate anti-tumor immunity? Well, fast forward to 2019, we now appreciate that there are both positive second signals that work together with the T cell receptor to stimulate T cell responses, but in addition, there are negative second signals that attenuate signals through the T cell receptor. What we've learned is that many of these inhibitory or negative second signals are actually key mediators of T cell T-cell tolerance that protect us against autoimmunity. And while this model for T-cell activation was initially proposed to explain how a naive T-cell would become activated, we now appreciate that these positive and negative second signals not only regulate naive T-cells but can control T-cells throughout their life history. They can control effector T cells, memory T cells, as well as these regulatory T cells. And so we now know that there are many T cell costimulatory pathways. And so this morning I'm going to focus on just two pathways, and I'm going to focus on these pathways because they serve as a paradigm for other pathways. And these are the ones that have translated to therapy. The first pathway that I'll talk about is the one that involves the B7-1 and B7-2 co-stimulatory molecules and their receptors CD28 and CTLA-4. And CTLA-4, of course, is the target of ipilimumab, the antibody that— it was the first co-inhibitory receptor that was translated to therapy. So first, let me introduce the players. The B7 molecules are potent costimulatory molecules. These molecules are expressed on antigen-presenting cells so that dendritic cells, macrophages, and B cells can express these molecules. The first molecule that was discovered was initially called B7. B7 and then cleverly renamed B71 when the second molecule was discovered by us in 1992. This molecule, B71, is also known as CD80. B72 is known as CD86. What distinguishes these molecules is the kinetics of their expression. When B71 was the only molecule that was known, it was a bit perplexing because it was expressed at low levels on resting antigen-presenting cells. And when you stimulated these cells, it could take several days for this molecule to be synthesized and upregulated. That didn't jive with the rapid need for expression of these co-stimulatory molecules to initiate an adaptive immune response. The second molecule that was discovered has these properties. Within hours in response to danger signals that can be lipopolysaccharide, a component of bacterial cell walls, for example, this molecule can be upregulated on dendritic cells, B cells, and macrophages, and then be poised to stimulate T cells. And so these molecules are closely related. They're members of the immunoglobulin gene family, have about 25% amino acid identity. As I mentioned, their expression can be enhanced by microbial products and also CD40-mediated activation. They have many overlapping functions, and one of the distinct features is the temporal kinetics of their expression. These molecules interact with two receptors, CD28 and CTLA-4. CD28 is constitutively expressed on all mouse T cells, on 95% of human CD4 cells, and about half of human CD8 cells. CTLA-4, in contrast, is upregulated on T cells upon their activation. It primarily is in the cytoplasm of cells, but upon the activation of these cells, it gets shuttled to the surface. CD28 and CTLA-4 have about 30% amino acid identity, and they have a locus, a component of their extracellular cellular domain referred to as the MYPPPY or the my puppy domain, which is what binds the B7 molecules. And CTLA-4 binds B7-1 and B7-2 at higher affinity than does CD28. So what do we know about CD28? And in this one slide I have summarized years of work, many graduate student and postdoctoral experience. We know that CD28 is the major co-stimulatory receptor for naive T cells, that when there is a dangerous signal and you get signal 1, the antigen presentation to the T cell receptor, that B7-2 becomes expressed and can engage CD28. As a result of signals through CD28 in concert with T cell receptor, there is an increase in cytokines, which are growth factors for these cells, such as IL-2. There is expression of growth factor receptors by the T cells, such as the IL-2 receptor. There is also elaboration of survival factors, such as Bcl-XL, and metabolism is stimulated. Glucose receptors are upregulated and glucose metabolism is upregulated. So these T cells then have the growth factors, the survival signals, and the metabolic signals, and the bioenergetics, so these cells then can proliferate and differentiate into effector cells that then can respond to the antigen. This understanding has translated to therapy, blockade of the B7-CD28 pathway. We first understood from animal models that you could block this pathway and prevent transplant rejection. Other studies Studies showed that you could prevent the induction and progression of autoimmune disease. The induction is of interest, but therapeutically preventing the progression of autoimmune disease has therapeutic significance. In addition, blockade of use of this pathway has been taking these B7 molecules and using them as adjuvants to promote vaccination. Genes for infectious diseases, and some of the earliest anti-tumor approaches were to put these B7 molecules into tumors to drive and stimulate the immune system. What is translated to therapy is a soluble form of CTLA-4, the high affinity receptor, and this is a fusion protein of the extracellular domain of CTLA-4 and the Fc tail of an IgZ molecule. So this is abatacept or Orencia that is used to treat rheumatoid arthritis. And this is also used in solid organ transplant. The idea here is that the abatacept or molecule is able to bind to the B7 molecules and compete so CD28 can no longer engage these B7 molecules and prevents the T-cell activation. So we'll turn to CTLA-4 now, which is the higher affinity receptor for B7-1 and B7-2, and it has been the mystery molecule in this pathway from its earliest days, it was debated because of its homology to CD28, was it also a stimulatory molecule or was it an inhibitory molecule? And some of the early studies with antibodies to CTLA-4 in vitro showed that if you had antibodies to CTLA-4, you could stimulate T-cell responses. And there was a debate as to whether these antibodies were engaging the receptor and stimulating responses. There were stimulatory responses to CD28 antibodies, so was that the case? Or were these antibodies blocking an inhibitory signal? There are ways that one can modulate antibodies to test these hypotheses, but about that time, my laboratory and Tak Mak's laboratory made a knockout mouse of CTLA-4 that really convinced the field that CTLA-4 had an inhibitory function. Now I'm biased, but I think this knockout has one of the most dramatic phenotypes. These animals die by 3 to 4 weeks of age. They develop massive splenomegaly and lymphadenopathy. There's spontaneous T-cell T-cell activation and multi-organ lymphocytic infiltrates. So this is the heart from one of these animals. All of these purple dots here are lymphocytes. These mice develop a giant cell myocarditis. They also develop a pancreatitis. Here you can no longer see the islets in the pancreas, and there's complete destruction of the pancreas. So this mouse showed us that these so-called co-stimulatory receptors could provide not only positive second signals, but also inhibitory second signals, and demonstrated that CTLA-4 had a critical role in downregulating T-cell responses. This phenotype looked for all the world like an autoimmune disease, and led us and others to ask whether CTLA-4 had a role in regulating tolerance and autoimmunity. And this pathway suddenly became much more complex because these B7 molecules could engage CD28 and promote T-cell activation, but when they engaged CTLA-4, they inhibited the immune response. So the outcome of this pathway is somehow a balance between CD28-mediated T-cell activation and CTLA-4-mediated inhibition. So what do we know about how CTLA-4 inhibits T-cell responses? And I'm going to divide this into two parts. The first is cell intrinsic mechanisms, meaning CTLA-4 has a function on T cells that express it. So we know, and I've already told you, that CTLA-4 is the high affinity receptor for the B7 molecules. So by virtue of this affinity, it could outcompete CD28 for binding these B7 molecules.. The other thing we know is the biophysical properties of CD28 and CTLA-4. The crystal structures for B7 binding to CTLA-4 were solved by an MD-PhD student, Jean-Claude Schwartz, when he was in Stan Nathanson's laboratory. And this is a cartoon where we have the T cell membrane going into the ceiling. And the antigen-presenting cell membrane going into the floor. And what we have here are the B7 molecules shown in blue as these B7 non-comatulent dimers and the CTLA-4 molecules shown in red. And it turns out that the CTLA-4 molecules can bend, and so they're able to bind two different B7 molecules.— and this leads to a very stable structure that's been referred to as a molecular zipper. Very few molecules do this. And so you can imagine here that there's signaling molecules that could align and lead to recruitment of signaling molecules and lead to this very stable structure. It turns out that when the crystal structure for CD28-B7 interactions was that CD28 cannot do this. It can only bind to one B7 molecule. So by virtue of affinity and biophysical properties, CTLA-4 could win in terms of interacting with these B7 molecules. In addition, we know that CTLA-4 intracellularly can compete for some of the same signaling molecules that associate with the cytoplasmic domain. But in addition, CTLA-4 is able to directly antagonize signals through the T cell receptor. So there's a bit of a yin-yang, whereas CD28 can promote entry into a cell cycle and IL-2 production, CTLA-4 inhibits entry into the cell cycle and inhibits IL-2 production. And T cell proliferation. But CTLA-4 is not only expressed on naive cells that become effector cells. CTLA-4 can be expressed on these regulatory T cells. These are cells that inhibit immune responses. And CTLA-4 is a direct transcriptional target of FOXP3, which is the cardinal transcriptional transcription factor that defines and expresses these suppressive regulatory T cells. So these regulatory T cells constitutively express high levels of CTLA-4. We know from studies in mouse models that CTLA-4 has a crucial role in regulatory T cell biology. If we knock out CTLA-4 in CTLA-4 only on regulatory T cells, these mice still will develop a fatal lymphoproliferative phenotype, but it's a bit delayed compared to the global deficiency of CTLA-4. So this indicates that CTLA-4 has an important role in Tregs, but Treg expression of CTLA-4 doesn't completely account for the phenotype of the CTLA-4 null mouse. And so we know that CTLA-4 is a mediator of T-reg function and really endows T-regs with suppressive capacity. So inhibitory functions of CTLA-4 have several types of functions. In addition, if we think about CTLA-4 cell extrinsic mechanisms influencing other types of cells, while regulatory T cells influence effector cells, so this could be a cell extrinsic mechanism. There also have been studies suggesting that when CTLA-4 binds to dendritic cells, there can be signaling into those dendritic cells through the B7 molecules that can cause them to acquire suppressive properties as well. In addition, we know that when Tregs make certain suppressive molecules, certain cytokines, this also can downmodulate B7 expression. So when we think about the mechanism of action of CTLA-4, these cell-intrinsic functions can directly block activation of T cells block signals through CD28, but in addition, CTLA-4 and regulatory T cells can influence antigen-presenting cells like dendritic cells by either signaling into these cells through the B7 molecules or through another mechanism, can actually pull the B7 molecules off the dendritic cells, a process known as transcytosis., and as a result of that, there are fewer B7 molecules that can engage CD28. So CTLA-4 can exert its functions in several ways. So what are the physiologic functions of CTLA-4 and its inhibitory functions? And so when we go back to the phenotype of the CTLA-4 knockout, that suggested that CTLA-4 may have a role in controlling tolerance and autoimmunity. And indeed, studies by my lab and others have shown that CTLA-4 does control regulatory T cells and self-reactive effector cells and regulate autoimmunity. In addition, there are polymorphisms in the CTLA-4 gene that are associated with human autoimmune diseases. Long before we had the sequencing of the entire genome, there were gene-wide association studies from Linda Wicker and John Todd that showed the polymorphisms in CTLA-4 were associated with human type 1 diabetes and familial Graves' disease. Now, CTLA-4 is a very compact gene. Gene. It has 4 exons and a domain that is a leader sequence. There is the IgV-like domain that can bind to B7 molecules on the cell surface, a transmembrane domain, and then a cytoplasmic signaling domain. There are also splice variants of CTLA-4, and it turns out the polymorphisms that are associated with human disease are associated with some of these splice variants, in particular the soluble form, and that's been more difficult and challenging to study. We still don't know what it does, but there's been very interesting work from Mike Lenardo's lab at the NIH that has defined several different diseases of CTLA-4, and I think there was a bit of a sense of humor here in that these diseases are referred to as CHAI and LATE. CHAI is an acronym for CTLA-4 haploinsufficiency with autoimmune infiltration. This is a disease, and you can see here different mutations throughout the CTLA-4 gene that can result in haploinsufficiency. These patients develop these manifestations later in childhood with variable penetrance. There can be enteropathy in the gut, there can be manifestations in the lung, as well as manifestations in the brain. The other disease, this LATE disease, is an acronym for LRBA deficiency with autism autoantibodies, Treg defects, autoimmune infiltration, and enteropathy. So here there is an alteration in a molecule known as LRBA, which is an LPS-responsive vesicle trafficking beige-like anchoring-containing gene. This is the protein that is key to shuttling CTLA-4 from its cytoplasmic location to the cell surface. And individuals who have this deficiency in LRBA can't get CTLA-4 to the cell surface, so it's another reason for CTLA-4 deficiency. This manifests early in childhood, and these individuals develop a variety of autoantibodies platelets to other immune cell components. They have defects in T-reg function and infiltration enteropathy, lung manifestations, as well as CNS manifestations as well. It turns out that CTLA-4 IgG was one of the treatments for this disease because it will block B7 molecules from engaging CD28. The other claim to fame, of course, for CTLA-4 is blocking CTLA-4 as a strategy to promote anti-tumor immunity. The idea here is that we know that CTLA-4 inhibits T-cell activation, so if we block CTLA-4, we then can have unopposed stimulation through through CD28 and promote an ongoing immune response. And it was Jim Allison who first had the idea when he saw the phenotype of the CTLA-4 knockout to develop this strategy. Early on, he was the one who had tried to stimulate immune responses by putting B7 molecules into tumors. That didn't work so well, but blockade of CTLA-4 in animal models showed dramatic effects. And this is just one example of an experiment where one is using and implanting a mouse melanoma, a B16 melanoma, and combining this with a melanoma vaccine and CTLA-4 blockade. So what you can see here is the way these These experiments are done, and they're done in a very humane fashion. All of these experiments are done with approval of an animal care and use committee. And although this says survival, these animals are sacrificed at humane endpoints if there's a certain amount, 20% weight loss, or the tumors get to a certain size. So here we're looking at what is deemed survival, but it's really these humane endpoints over time. So you can see that if you implant these tumors and give a control hamster antibody, that these animals all die. If you give anti-CTLA-4 alone, about half of these animals, 40%, will survive for a long period of time. If you give the vaccine alone, you can have some benefit. But the exciting part was by this combination, you could see that there was a long-term benefit in these animals. And so this work showed that CTLA-4 blockade had the potential to enhance T cell responses and lead to tumor clearance in animal models that could be accompanied by long-lasting immunity. However, even in these studies, it was appreciated— the good news was that the tumors were cleared, but around these sites of vaccination, there was depigmentation, vitiligo, an immune response against the melanocytes that you could see developing over time, and this could become generalized. So the good news was you could lead to anti-tumor immunity, but there could be the potential for autoimmunity. And of course, This was translated to therapy with ipilimumab, and Steve Hodi at the Dana-Farber Cancer Institute down the street led the clinical trial that led to the approval of ipilimumab for melanoma. And this really illustrated the potential. The 20% of patients who had a long-term response first illustrated the potential to induce long-lasting responses and initiated the first excitement that you could produce a tail on the survival curve. But this was associated with toxicities, in particular colitis, but did lead to the FDA approval. So to summarize then, CTLA-4 was the first co-inhibitory receptor discovered and translated to therapy. The fatal inflammatory phenotype of the of CTLA-4 knockout revealed the critical inhibitory function of CTLA-4 and led to further work that demonstrated that CTLA-4 is a key mediator of T cell tolerance and prevention of autoimmunity. Blockade of this inhibitory receptor can promote anti-tumor immunity known as checkpoint blockade and stimulate the anti-tumor immunity by reducing inhibitory signals. CTLA-4 blockade was approved for melanoma in 2011 and really is what catalyzed the excitement in this field for checkpoint blockade. So to summarize what I've told you about this pathway, the B7-CD28 interaction is key for the initial activation of naive T cells. T cells. This knowledge is translated to therapy with CTLA-4 Ig, a soluble form of CTLA-4 that can be used in the setting of rheumatoid arthritis and transplant rejection to block T cell activation and affect your T cell responses. B7-CTLA-4 interactions are critical for inhibiting immune responses. And are key mediators of T cell tolerance. And blockade of this pathway, of course, is critical for stimulating anti-tumor immunity by reducing inhibitory signals. But this is really just the beginning. There are now multiple inhibitory pathways that are known to regulate this balance for T cell activation and tolerance. Now, I have several versions of this slide. I have this version that you can read, and I have another version that lists all of the inhibitory pathways that have been discovered. And one of the fascinating aspects of immunology is how many inhibitory pathways there are. And these— many of these pathways regulate this balance between stimulating T cells and mediating tolerance. And many of these molecules are now druggable targets for tumor immunotherapy and have translated to therapy. So CTLA-4 was the first, the PD-1 pathway is the second, and that's the pathway I'm going to focus on for the remainder of my talk. But there are other pathways as well, such as TIM-3, LAG-3, PD-1, PD-3, TIGIT, CD160, and others. And we'll come back to those toward the end of my talk. I'm going to focus on PD-1 since this is the other pathway that's translated to therapy. And I'll use this cartoon to introduce this pathway. The PD-1 receptor is also like CTLA-4 upregulated on T cells upon their activation. PD-1 has two ligands, PD-L1 and PD-L2, and when they engage PD-1, PD-1 becomes phosphorylated on tyrosine motifs in its cytoplasmic domain. This leads to the association of protein tyrosine phosphatases such as SHP2, which then can lead to dephosphorylation of kinases downstream of the T T-cell receptor or downstream of CD28. As a result, there's reduced signaling through the T-cell receptor in CD28 and an inhibition of T-cell responses, inhibition of effector T-cell responses, effector cytokine production, as well as cytolytic T-cell responses. Let me tell you a little about the ligands for PD-1. PD-L1 and PD-L2. My lab, together with Gordon Freeman's lab, was involved with discovering PD-L1 and PD-L2 and showing that they bound to PD-1, thereby defining this pathway. One of the things that fascinated us from the beginning were the expression patterns of PD-L1 and PD-L2. PD-L1 can be expressed on a variety of hematopoietic immunopoietic cells as shown here, dendritic cells, macrophages, B cells, and T cells. But in addition, we found that PD-L1 could be expressed on many other cell types, on vascular endothelium that lines the blood vessels, a variety of types of epithelia, muscle cells, cells in the liver, cells in the islets in the pancreas, sites of so-called immune privilege. PD-L1 can be expressed on syncytiotrophoblasts, in the placenta, and cells in the eye. This expression pattern suggested to us that perhaps PD-L1 could control responses locally within tissue microenvironments. PD-L2, in contrast, is more restricted in its expression. It's expressed mainly on hematopoietic hematopoietic cells, but now we know that it can be expressed on some non-hematopoietic cells, in particular cells in the lung microenvironment. This pathway seems to be very important for respiratory tolerance. Interferons alpha, beta, and gamma are very potent stimuli for upregulating PD-L1. These cytokines can also upregulate PD-L2, though IL-4 and GM-CSF are the more potent stimuli. So you can think about PD-L1 and PD-L2, and this pathway is an important feedback mechanism. After you get T-cell activation, the cytokines of adaptive immunity are expressed and then downregulate— upregulate PD-L1 and PD-L2 so they can then inhibit immune response. Responses. And when one thinks about this pathway, it's a poster child for what inhibitory pathways can do in the immune system. As I already mentioned, PD-1 inhibitory signals counterbalance positive signals through the T-cell receptor and CD28 co-stimulatory signals. This pathway also has an important role in regulating immune tolerance. This pathway controls not only the induction of tolerance but also the maintenance of tolerance. We found that PD-L1 in non-hematopoietic cells has an important role in protecting the tissues. You can think about PD-L1 as P means protect. PD-L1 is really a guardian of tissue health. It can help control resolution of inflammation. So you don't get immune-mediated tissue damage. And it also can shield target organs from autoimmune attack. So when we think about an immune response in the PD-1 pathway, this pathway can control tolerance in several ways. It can prevent the initial activation of self-reactive CD4 or CD8 cells. And in addition can control potentially pathogenic CD4 and CD8 cells. And we found that PD-L1 on the non-hematopoietic cells can have a role. PD-L1 can be expressed on vascular endothelium, and there's the potential for PD-L1 to control the extravasation of T cells into tissues, areas of current research. But it's the smartest microbes, the tumors and the smartest immunologists, the tumors and the microbes that have over and over again exploited this pathway to their own benefit to inhibit immune responses to prevent their eradication. This pathway contributes to a process of T-cell dysfunction known as T-cell exhaustion that develops during chronic antigen stimulation. So to illustrate this, we'll compare an acute viral infection like influenza to a chronic viral infection such as HIV, and there's a mouse model of this, that during an acute infection you will get activation of CD8 cells, for example, that, that can differentiate into killer cells that can eradicate the virus, and then you have a transition to memory. What happens in the setting of a chronic infection is that there is a chronic stimulation of the immune system. As a result, these cells progressively lose their function. First, they lose the ability to make IL-2, then they lose the ability to kill and to make other cytokines. We now appreciate that there are multiple subsets sets of these cells, some that are more stem-like and some that just fade away. In cancer, we know that there's a similar process that develops. We know that PD-1 can be highly upregulated during an acute infection, but when the infection is cleared and there's no longer a TCR stimulus, PD-1 expression goes away.. But in the setting of a chronic infection in a mouse model, PD-1 stays high. And work that we did in collaboration with Rafi Ahmed's lab, John Wherry, Dan Barber, and Gordon Freeman showed that the cells that highly expressed PD-1 don't function well. And importantly, when you blocked this pathway in a chronic infection model in mice, you could revive these exhausted infected cells, and now they could kill. Within the same year, work from Bruce Walker's lab and Rafik Sekaly's lab showed that this was also the case for human chronic infections. What I want to point out here is that PD-1 can be expressed upon T-cell activation or when there are exhausted dysfunctional T-cells. So PD-1 expression alone you cannot determine whether these cells have become recently activated or whether these cells are exhausted. Further work is needed. And these tumor-infiltrating lymphocytes behave like these exhausted cells. But what's special about the tumors is that they themselves can also express the ligands for PD-1. When we first reported PD-L2 as a second binding partner for PD-1, PD-1, we also showed that PD-L1 could be expressed on certain cancer cell lines. And there was an article that was in the Dana-Farber Cancer Newsletter, which also seems prophetic at this point, that described PD-L1 as cancer's shield against the immune system. And we now know that PD-L1 can be expressed on the surface of about 30% 30% of solid tumors and certain hematologic malignancies, and that PD-L1 on tumor cells themselves can contribute to inhibiting anti-tumor immunity. So in the setting of cancer then, we have these CD8 cells that highly express PD-1 and they're dysfunctional. PD-L1 can be expressed on the tumor cell, on antigen-presenting presenting cells in the tumor microenvironment, as well as endothelial cells and other stromal cells. And the antibody drug that's been developed, antibodies to PD-1 or PD-L1, block this inhibitory signal by blocking this interaction and can unleash a potent anti-tumor immune response, leading to increased cytokine production by these cells, such as interferon gamma, as well as increased killing. And as you know, this has translated to therapy with impressive success. PD-1 or PD-L1 inhibitors have now been approved by the FDA for over 14 different tumor indications, and this has stimulated great excitement in the field. I know that you'll have more talks about this But I just want to mention some of the key issues in the field. As you may have noticed, there's a broad range of responsiveness. Some patients respond better than others. Some tumors respond better. And so we work in the field now is trying to identify who will respond to PD-1. Can we identify biomarkers for that? Work under active control, active investigation. We know that PD-L1 expression alone is not good enough. It will likely be a combination of a variety of transcriptional signatures, PD-L1 expression, as well as perhaps the microbiota. Also, we know that there are some patients who never respond. How can we identify those patients? Some patients initially respond and become resistant. What are those mechanisms? We need to understand this to develop rational combination therapies for those who don't respond to PD-1. In addition, what's been mentioned already are these immune-mediated adverse events. These have been appreciated more recently with the FDA approval, but similarly, questions about how to identify who is at risk and how can we dissociate the adverse events with promoting anti-tumor immunity? One area under active investigation are polymorphisms in CTLA-4 and PD-1, for example. Can they help prevent— predict the risk, the polymorphisms that are associated with autoimmunity? But T-cell exhaustion is more than PD-1. Many of those other inhibitors inhibitory receptors that I mentioned to you can be co-expressed on these exhausted cells with PD-1. So PD-1 plus LAG-3 or TIM-3 or TIGIT. And studies have shown that if we look at the number of inhibitory receptors that are expressed and we plot that against functionality here, poly-functionality cytokine production, you can see that the more inhibitory receptors that are expressed, the less functional these cells become. So that led to studies first in animal models saying, well, if these cells co-express PD-1 and another inhibitory receptor, what happens if we block more than one? And it's been found that if you co-block PD-1 and many other of these receptors, you can have a better benefit than blocking PD-1 alone. But only PD-1 blockade by itself has substantial benefits. So the future is in combination therapy, and in closing here, I'd just like to make a few points. The first is that we really need to know more about these pathways to rationally learn how to combine PD-1, which has become a building block for combination therapies, with other strategies. For example, one strategy is to block PD-1 and these other inhibitory receptors. PD-1 and CTLA-4 co-blockade is FDA approved. PD-1 blockade with LAG-3, TIGIT-TIM-3, or CD-47 are all in clinical trials. But we have no idea whether these co-blockades are acting on the same cell types or the same molecular circuits, work we need to do in order to understand how to combine things rationally. Other approaches are to block PD-1 and then stimulate the immune system in some way. Agonizing stimulatory receptors using IL-2 or different components of the innate immune system. Blocking PD-1 and giving a kinase inhibitor or chemotherapies. Some of this has also translated to FDA approval, PD-1 blockade and cisplatin. But we're also realizing that these types of therapies have effects on the immune system and sometimes on different cell types. So we need to understand that better in order to understand how to most effectively combine these therapies. Other approaches are to combine PD-1 blockade with angiogenesis blockade, radiation, or some of these epigenetic modifiers. We know that Exhausted T cells have a different epigenetic state that PD-1 blockade does not erase. So epigenetic modifiers are a rational type of combination, and in animal models have shown benefit. Other approaches are to think about different tumor microenvironments, so-called cold tumors versus hot, where we need to bring those T cells into these immune deserts that don't have T T cells. And so these types of strategies involve vaccines, oncolytic viruses, or engineered T cells, and understanding how to combine this with PD-1, or also PD-1 with the microbiome. So to summarize then, the success in checkpoint blockade has led to a paradigm shift in cancer therapy. We're appreciating that tumor immunity is limited by physiological immunoregulatory mechanisms. And the future is in combination therapies that we hope will extend the benefit. But we need a better understanding of the mechanisms that control anti-tumor immunity. We need to understand the tumor microenvironment deeply to understand how host factors such as age, gender, obesity, and the microbiome impact anti-tumor immunity and therapy. Therapy, and to integrate an understanding of basic biology into the next generation of therapies. And this is going to come, I think, critically from detailed evaluation of patients to understand mechanisms of response, resistance, and adverse events. And in this way, we'll be able to understand how to increase the benefit and mitigate the risks. So I'll stop at at this point, and I'm happy to answer questions if there's time. Maybe we'll take one or two questions. Do you want to come to the mic?

Hi, Dr. Shah. You mentioned that CTLA-4 expression on Tregs can limit their activation and proliferation. So when we use an anti-CTLA antibody, on one hand it can remove the brakes off the CD8-positive T cells. On the other hand, can it increase the T regulatory cell activation and limit the activation of the CD8-positive T cells?

Yes, so CTLA-4 blockade can do two things. It can increase the function of the CD8 cells directly, but it also can block the function of regulatory T cells so they don't work so well. So as a result of reducing that suppressive cell type from working, there can be further activation of these effector cell responses. Responses. So it can do two types— it can work in two ways: reduce Treg function and increase effector T cell responses that can kill tumor cells. So it can reduce Treg function, is that right? Yes, so it can impair regulatory T cell function. These cells don't work well when you block CTLA-4.

But you mentioned that CTLA-4 expression on Tregs can limit Treg activation and proliferation.

That's right, yes. Okay, but when you block it, you— it turns out that you'll get more of them, but they don't work well. Okay, okay, thank you so much. Okay, thank you so much.